Description: This paper is a component of the doctoral thesis "Discourses of school leadership in Norway". The paper illuminates political descriptions of school leadership based on interviews with key politicians. The main goal is to describe the political discourses of school leadership based on the political oral language.Earlier research shows that studies of written educational policy only give a partial picture of the existing discourses of school leadership on policy level (Valle, 2005). The main reason is the objectiveness of policy documents which not compromises doubt, difficult subjects at the sacrifice of unassailable rhetoric. The study is based on interviews with seven members from the following political parties: The Labour Party, The Progress Party, The Conservative Party, The Socialist Left Party, The Christian Democratic Party, The Centre Party and The Liberals. All informants were main responsible for the basis instruction of the primary school in their committee work.The analysis is based on two research questions: First, How do politicians describe successful school leadership? And second: What are the main differences between the political parties? Methodology: The theoretical perspectives are embedded in discourse theory (Laclau & Mouffe, 1997). Discourse analysis focuses on patterns of speech. This method of analysis offers a perspective in social life that contains both methodological and conceptual elements. Discourse analysis involves ways of thinking about discourse and ways of treating discourse as data. Discourse analysis is thus not simply an alternative to conventional methodologies; it is an alternative to the perspectives in which those methodologies are embedded. There are multiple definitions of discourse and of what counts as discourse (e.g., spoken language, written language, language use above the level of the sentence, etc.). Wood (2000: 3) claims that discourse may not just be seen as an object, but as a way of treating language. In other words, language is taken not to be simply a tool for description and a medium of communication, but as social practice, as a way of doing things. From this perspective, discourse analysis will be suitable for data material like policy documents and interviews with politicians. Such data can be treated not only as descriptions of a certain topic, but also as a way of social practice. Conclusions: The preliminary findings can be described like this: First, The oral language is an important source of knowledge in understanding education policy. In addition, qualitative interviews can provide useful data in such understanding. Second, Norwegian politicians describe school leadership in different ways. The descriptions of the main goal for school activity seem to diverge. Political reasoning for change in educational policy also diverges among the political parties. Third, oral language leads to other descriptions of school leadership according to policy documents and argumentation through media.Finally, there is a tension between individual oriented school leadership and collective oriented school leadership.